Methods I Lab Lecture 3 Homework submissions Name

Methods I Lab Lecture 3

Homework submissions Name files using the following format: HW#_Last. Name. doc or HW#Last. Name. pdf Examples: HW 2_Muhammad. doc HW 4_Hyer. pdf If you have the option to produce a word document, let that be the default choice. If you’ve missed something and we would like to comment, it is far easier for us and you if it is a word document

Homework #1 Common Mistakes Please be careful to answer all questions and every part of each question Graphs should be stand-alone The information should be all inclusive If someone cannot look at the graph completely out of context and know exactly what they are looking at, reconsider your titling, description, and labeling

Homework #1 Common Mistakes NOTE: Standard deviation is not guaranteed to reduce when N increases Smaller sample sizes only increase the probability and size of error. There is not promise that you’ve over or under estimated in the first place.

N = 25 N = 100 N = 250 • Suppose you knew the ‘true’ mean and standard deviation of a population… • Would every subset of that population reflect those values? • Here, subsets of size 25, 100, and 250 are taken from a population of size 2, 000 with a mean of 0 and SD of 15

• There’s no guarantee that a subset of the population will reflect the ‘true’ population parameters • In 1908, W. S. Gossett introduced the idea that there is uncertainty when estimating parameters of a small sample • Here, we see that outliers tended to have a smaller sample size than those mostly centered around the true parameter values

Onto the good stuff…

Data in SPSS - Exercise Make an ‘ID’ variable and list numbers 1 to 20 Generate 2 random numbers using a uniform distribution Use Rand 1 and Rand 2 for variable names Transform Compute Variable Function Group: Random Number Functions &…: Rv. Uniform Min: 0 Max: 1

Data in SPSS - Exercise Generate another number that is the product of your first two I’m naming mine Rand 1 and 2 Transform Compute Variable Rand 1*Rand 2

Data in SPSS - Exercise Make a categorical variable for each of the variables Have the first be into three groups, the second into two groups, and the third into two groups In this example Rand 1 cutpoints: . 25 and. 75 Rand 2 cutpoint: . 75 Rand 1 and 2 cutpoint: . 5

Analysis in SPSS - Exercise Produce crosstab tables between the first two random variables and the variable you Computed Analyze Descriptive Statistics Crosstabs… Row(s) : Column(s):

Analysis in SPSS - Exercise Rename your two binary variables to something that indicates positive disease status for one and positive disease test Doesn’t matter which is which

Analysis in SPSS - Exercise Produce crosstabs with disease status as the column variable and disease testing status on the row Under ‘Cells…’, request ‘Row’, ‘Column’, and ‘Total’ percentages Analyze Descriptive Statistics Crosstabs… Row(s) : Positive Test Column(s): Positive Disease Cells… Row, Column, and Total Percentages

Analysis in SPSS - Exercise From the output, identify the positive and negative predictive values Consider what the PPV and NPV is and identify the corresponding % You shouldn’t get the exact same results as the example without the exact same dataset If you would like to practice with this dataset, email Madison or Lutfiyya and request it